Planet Review
Planet Review is an evidence-first workflow for reviewing transit-like signals without turning every periodic peak into a planet claim.
The problem
Transit-search tools can produce convincing peaks, but learners and early reviewers may not know when a result is a known signal, a harmonic, an artifact, or simply inconclusive. Planet Review organizes that evidence into a cautious, reproducible review instead of rewarding the most exciting interpretation.
Who it is for
The Build Week version is designed for astronomy students, educators, citizen-science learners, and technical reviewers who want a transparent example of how to separate evidence, limitations, and follow-up recommendations.
What the demo shows
- A deterministic offline Kepler-265 benchmark workflow
- Recognition of the 34.057671-day peak as part of Kepler-265 c's harmonic family rather than a new planet
- Retirement of the earlier 12.160120-day residual as inconclusive under stricter cadence and event-consistency requirements
- A clearly labeled synthetic software-test fixture that is not presented as telescope evidence
- A Streamlit interface, command-line workflow, Markdown and JSON reporting, automated tests, and GitHub Actions
How GPT-5.6 and Codex were used
GPT-5.6 helped challenge assumptions, shape the architecture, define scientific claim controls, identify circular or overstated interpretations, and decide what evidence should block promotion. Codex implemented the typed analysis modules, deterministic benchmark pipeline, Streamlit interface, report generation, command-line path, documentation, repository safeguards, tests, and CI workflow. Final scientific and product decisions were reviewed by the entrant before merging.
Technical implementation
The same typed Kepler-265 pipeline is shared by the UI, reports, JSON export, CLI, and tests. The repository includes conservative classification guardrails, harmonic-family checks, preserved historical dispositions, explicit limitations, and 46 passing tests.
Judge setup
Use Python 3.11. From a fresh clone:
python -m venv .venv
python -m pip install --upgrade pip
python -m pip install -r requirements.txt
python -m pip install -e .
pytest -q
streamlit run app/streamlit_app.py
No account, paid service, credential, external hardware, or live observational download is required. The app opens directly to the offline Kepler-265 benchmark.
Limits
Planet Review does not confirm planets, calculate validated planet probabilities, reproduce a fresh live Kepler search, or replace professional astronomical vetting. Its purpose is to organize evidence, warnings, limitations, and recommended follow-up in a cautious and reproducible way.
Log in or sign up for Devpost to join the conversation.